Submitted by Anna Schnell
on
In March, Professor Emily M. Bender participated in an event titled "The Great Chatbot Debate," hosted by the Computer History Museum (Mountain View, CA) and co-organized by IEEE Spectrum.
In the debate moderated by IEEE's Eliza Strickland, Bender went up against OpenAI's Dr. Sébastien Bubeck over the question "Do LLMs Understand?"
Bender highlighted the relevance of linguistics to understanding both how LLMs work and why their illusion of understanding is so powerful to make the case that no, they emphatically do not understand. The full debate can be viewed here, on CHM's YouTube channel.